I approached the article by Alan et al1 with great exceptions. Here at last was an example of the long-term follow-up of perinatal interventions that cerebral palsy (CP) epidemiologists have been calling for: a call supported by the results of this article, in which the results of short-term follow-up are not mirrored by the long-term follow-up. I was disappointed.
The article is largely devoted to abnormalities of ultrasound images in the neonatal period: the trial of indomethacin promised in the title is almost completely ignored. It has already been frequently reported that these ultrasound abnormalities (USAs) are associated, but only associated, with CP. The false-positive rate is always too high for confident prediction. Table 6 would have indicated this if the false-positive rate had been correctly calculated as it was defined on page 581: the proportion (although I presume that it is given in the table as a percentage)